Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

KT120 tests, compared to 6550.

795 views
Skip to first unread message

Patrick Turner

unread,
Feb 5, 2012, 7:31:16 AM2/5/12
to
Over the last fortnight I rebuilt my benchtop power supply to be able
to test KT120 with a regulated 600Vdc anode supply, at up to 320mA. I
placed shunt regulation for the screens on the breadboard for the pair
of KT120. Fixed bias was used, with a balance pot to equalise Ia in
each tube. Operation in all tests were limited to class AB1 with no
grid current.
The OPT is a 8Kg monster with negligible winding resistance, and lots
of interleaving, so its character does not affect the function of the
tubes. The OPT has 20% CFB, so that where there is 200Vrms+ at an
anode, there is 50Vrms- at the cathode. Depending on load, applied
grid signal is about 60Vrms at middle load values.

OPT ZR = 1,233:1, TR = 35.1:1.

The drive signal maximum for test1 below were the highest, and up to
115Vrms was needed to be applied to each KT120 grid. This was obtained
by using another tube amp and taking out the anode to anode signal via
10uF caps, and applied to quite high value biasing R. The maximum
available drive was 276Vrms from each 6CA7 anode from an ultralinear
power amp with 17dB global FB, so THD in the drive signal < 0.1% at
276Vrms.
At the levels of 115Vrms used, THD was considerably lower than for
276Vrms. Using an OPT with 20% CFB meant that the distortion in all
the tests remained low, and the Russian KT120 exhibited no nasty
surprises when compared to average condition Russian 6550 used to
explore the differences between the two tube types.

KT120 test1 :-
Ea = +600Vdc, Eg2 = +400Vdc and Idle Ia = 50mA, idle Ig2 was 4mA per
tube. Fixed Eg1 at -47Vdc. At absolute maximum PO possible, Ig2 rose
up to 17mA per tube.
Po versus RLa-a as follows, at clipping onset :-
2k0 = 60W, 4k0 = 102W, 4k8 = 115W, 6k0 = 102W, 8k0 = 80W, 12k0 = 56W,
16k0 = 45W.

KT120 test2 :-
Ea = +500Vdc, Eg2 = +500Vdc, Idle = 50mA, Ig2 slightly higher than in
test1.
Max Ig2 rose to 22mA per tube at absolute max possible Po.
Po versus RLa-a as follows, at clipping onset :-
1k0 = 90W, 2k0 = 127W, 2k5 = 130W, 4k0 = 100W, 4k8 = 85W, 6k0 = 72W,
8k0 = 58W, 12k0 = 40W, 16k0 = 27W.

Conclusions.
The effect of having the higher screen voltage in test2 raises the
knee of the diode curve. The diode curve in fact is the line for Ra
when Eg1 = 0V. With Eg2 = +400V, the curve is a straight line from
zero to 50V x 350mA, then the line rolls over to the left for 100V x
500mA, and 150V x 550mA. The knee shape could be lower for where class
A is intended, so Eg2 could be lower. But for highest class AB Po with
the lowest RLa-a load, the Eg2 should be ket as high as possible, and
well regulated. With Ea at +600V, I recommend Eg2 not be higher than
+450V, which would increase max Po possible to about 125W.

With Eg2 = +500V, then the diode line is nearly straight to 100V x
650mA, before rolling over at the knee to 200V x 760mA. The extra
+100V of Eg2 gives the tube a much higher max Ia capability with low
load values.

The curves I plotted for test1 and test2 show that test1 has the
higher spread of power, 80W or more is available between 3k2 and 7k9.

Test2 shows 80W or more of power is available from 1k0 to 5k3.

6550, Test3 :-
The 6550 was not tested wth Ea = +600V and Eg2 = +400V, as it is felt
by this technician that reliability can suffer with such high Ea for
6550.
But it was tested with Ea = Eg2 = 500Vdc.
1k25 = 90W, 2k5 = 112W, 3k7 = 93W, 4k9 = 77W, 7k3 = 57W, 9k8 = 45W,
12k0 = 32W, 16k0 = 24W

Conclusion. The results show that rather a small amount of extra Po is
available if the KT120 is used to replace 6550, or KT88 as currently
produced.

However, the tendency of the KT120 to over heat with abuse is much
less than for 6550, so that use with Ea at +600V and Eg2 at +450V
would be OK.

I fond that a KT120 tended to draw less Ia at idle if plugged into a
tube socket which previously had 6550, so it would appear it is safe
to plug in the KT120 to replace all other beam tetrodes, ( although
please don't assume anything, and don't assume KT12 will be OK where
there previously were EL34 or 6CA7. )

The distortion character witnessed in all tests was unremarkable, and
what I expected when a substantial amount of CFB is used, 20%, in this
case. The tube electrode signal relationships with 20% CFB is the same
as for UL with 20% taps.

I tested output resistance of the output stage as follows :-
Set up amp with 10.0Vrms, no load connected. The connect 4 ohms,
without changing anything else. The Rout = Vo drop / current in 4
ohms.
In this case, with KT120, Rout = 1.0 ohms, which includes maybe 0.1
ohms of Rw at the secondary. With ZR = 1,233, then Ra-a is effectively
reduced to 1,233 ohms, or Ra of one tube is around 600 ohms.

With the same test performed on 6550 in the same circuit and op
conditions, the Rout = 1.5 ohms, giving effective Ra-a = 1,850 ohms,
or around 900 ohms per tube.
So the KT120 must have a higher figure of merit, ie gm and µ is
higher, so Ra must be lower.

In conclusion, the KT120 performs rather better than is indicated by
the attrocious data sheets currently available from the makers, drawn
up by lazy bastards who seem to have forgotten what it takes to really
test a tube properly, and to provide curves without lies.

The KT120 is a slight betterment on the KT90EH which I like a lot. For
those NOT interested in trying to force huge amounts of class AB1 Po
from any tubes they lay their hands on, the KT120 allows those
interested in pure class A to have Pda at a safe 40W per tube
continuously, thus allowing up to 17W of class A1 Po, or 34W from a
pair. I have seen ppl try to run KT88 and 6550 at Pda = 40W, and seen
how the tubes didn't last long.
But with the higher 60W Pda rating, 40W should be OK.

I have not tested KT120 for triode connection, but I suspect they
should be good. 12Watts in SET may be easily possible from 1 tube, ie,
about twice what one gets from a single 6CA7 or EL34.

The simplest way to use KT120 for best fidelity and for home hi-fi use
is to use the UL connection because CFB OPTs are very rare off the
shelf. The KT120 will very much suit use in amps such as my 5050 at
http://www.turneraudio.com.au/Integrated5050.htm

Everyone knows huge Po is possible from PP KT120, but what audio
enthusiasts was is finesse, which means the first 30W is the most
important. I think they may find KT120 a fine replacement for all
previous octal beam tetrodes.

Patrick Turner.

rrus...@hotmail.com

unread,
Feb 5, 2012, 7:49:06 PM2/5/12
to
> shelf. The KT120 will very much suit use in amps such as my 5050 athttp://www.turneraudio.com.au/Integrated5050.htm
>
> Everyone knows huge Po is possible from PP KT120, but what audio
> enthusiasts was is finesse, which means the first 30W is the most
> important. I think they may find KT120 a fine replacement for all
> previous octal beam tetrodes.

Why produce a new beam power tetrode/pentode when the hi-fi market of
profitability is triodes? That is where the money is because of
pricing.

High end audio has access to beam power tubes because of guitar amps,
and they don't need any bigger tube than they now have.

Patrick Turner

unread,
Feb 6, 2012, 5:04:05 AM2/6/12
to
On Feb 6, 11:49 am, rruss...@hotmail.com wrote:

snip for brevity...

>  Why produce a new beam power tetrode/pentode when the hi-fi market of
> profitability is triodes? That is where the money is because of
> pricing.

>  High end audio has access to beam power tubes because of guitar amps,
> and they don't need any bigger tube than they now have.

Well, I'll probably have to let you consider a couple of things.
The reason for much of Russian tube production is that there are at
least 100 tube guitar amps for every hi-fi tube amp in the world.
Musos like as much bang for buck as can be cheaply afforded, so if
someone comes up with a way of making a better beam tet than has
existed before now, then its a winner, even if there were no takers
for triodes at all.

I have not tried KT90 strapped as a triode, but methinks it'd have to
be awfully good compared to using a 300B.
Any NEW beam tet usally means another good triode if that is wanted.

Having a *bigger* KT120 with rated Pda of 60W means that there *IS NO*
problem using just a pair where before you'd have used a quad of EL34,
6CA7, 6L6, KT66, etc. Audiophiles will, I think, embrace the KT120
favourably because the higher Pda allows for higher class A operation.
ARC are using them already. For all I know, maybe one could add yet
another 12mm to the anode height and make a tube with Pda = 72W, but
probably it won't be done because it may not fit under covers over
tubes in existing amps fitted with 6550/KT88, etc.

So, in conclusion, I think Mike Mathews deserves a gong for
contributing what appears to be another very fine tube to the list of
those already being produced.

In about 2005, I fitted KT90 to a pair of Quad-II amps I re-engineered
and all operating measurements improved. They were set up to draw
slightly less Ia than the original KT66 which laboured hard in those
amps. Ea was kept the same. So, KT90 used like that last a heck of a
long time, and give better sound. I've also used KT88 in Quad-II.
Maybe KT120 would be *over the top*. One reason is that heater current
for KT120 is 10% more than KT88, and this frightens ppl into not using
them. But KT88 work OK in Quad-II because the amps are made to power
old fashioned AM and FM tuners etc and a No 22 preamp unit. I fitted
KT88 to a Quad-II amp in 1998, and its never missed a beat since,
despite the hot weather in Cooma summers. But then I also reduced Ia,
junked the GZ32 rectifier entirely, used fixed bias, and drilled lotsa
holes in the bottom plate and re-set the sub-plate holding KT88
sockets to allow an air flow up through the amp, so, it deals with its
waste heat better because there is less of it and improved ventilation
that Peter Walker never considered.

In about 2002, I emailed the EI factory in Yugoslavia, with a request
that they make a PT40, ie, a Patrick Triode 40Watt tube with an OCTAL
base, and 6.3V indirectly heated cathode, so that its a drop in
replacement for most octal beam and pentode tubes, except it'd be a
REAL triode without a screen. They already had been making their own
version of KT90, and making a triode with an anode with same internal
dia as the screen helix so that KT90 bits and pieces could be used.
Well, their response came weeks later, and it seemed the guy was
drunk. It was also after the Kosovo war, and the EI factory became
moribund aftewards because many staff were needed to rebuild th
country, or were generally so pissed off they took to the bottle. They
also said how they'd been trying to develop exactly what I said
audiphiles wanted. There were a couple more emails, then just bullshit
from them, and I guess there just wasn't any money to develop a new
tube, and no real incentive, and probably much mental depression, even
though they had been making some good tubes. Companies should never
sit on their laurels, they need to get off their arses and inovate,
and improve, and not stagnate, let alone go out under a fog of vodka
and negativity. So the *OLD* experts have to stay with a spring in
their step and retain the freshness of youth but without being
youthful fools.

Jus' my 3.14 cents whirth.

Patrick Turner.

Roger Jones

unread,
Feb 6, 2012, 6:42:28 PM2/6/12
to
On Feb 5, 7:31 am, Patrick Turner <i...@turneraudio.com.au> wrote:

(SNIP)

> The drive signal maximum for test1 below were the highest, and up to
> 115Vrms was needed to be applied to each KT120 grid. This was obtained
> by using another tube amp and taking out the anode to anode signal via
> 10uF caps, and applied to quite high value biasing R. The maximum
> available drive was 276Vrms from each 6CA7 anode from an ultralinear
> power amp with 17dB global FB, so THD in the drive signal < 0.1% at
> 276Vrms.

(SNIP).

> Patrick Turner.

Patrick,
Just one comment on the above driver test rig... I have been bread-
boarding a pair of 7534 tubes with a view to making a more than "30
watt amplifier" (not the subject of this post.) To get a low enough
impedance drive at a high enough voltage, I used a smallish P-P tube
OPT, reversed, driven by a separate 10 watt audio amplifier from its 8
ohm tap to the small OPT's speaker side; amplifier terminated by 15
ohms across 15 ohm tap to keep it happy. A lot of drive volts were
available on the OPT "plate side"... I did not measure them but I
eventually had the pair of 7534 tubes giving 60 watts RMS (via a big
Hammond OPT) at 400 Hz at clipping, with the grid drive side
undistorted on the 'scope (class AB1 7534's cathode biased on 500 VDC B
+, but that's another story.)
This way you get to keep your fingers out of the driver amplifier's HV
section! Of course, you can even use a spare s/s receiver amplifier
as a driver unit.
Cheers,
Roger

rrus...@hotmail.com

unread,
Feb 6, 2012, 10:19:07 PM2/6/12
to
I am not saying this is a bad tube, just that production resources
could be more profitably aimed at making triodes rather than a beam
power tube that guitar amps probably won't much use. The trend is
smaller tubes.

Even at that I have long advocated a BASS amp with a single pair of
transmitting tubes, the Svet SV811-x lo mu single ended 811 mutants
would work great. Then again so would some pentodes or tetrodes, the
4-65A is much underused as is the 4E27. The 813 is too physically big.

Patrick Turner

unread,
Feb 8, 2012, 10:45:03 PM2/8/12
to
The world mainly does not share your views about what is/isn't a waste
of resources or what constitutes profitability. My forecast is that
many people will vote with their wallet with KT120. And if you want a
nice fat triode, connect the screen to anode.


>
>  Even at that I have long advocated a BASS amp with a single pair of
> transmitting tubes, the Svet SV811-x lo mu single ended 811 mutants

Yeah, ppl everywhere COULD have used different tubes than what you
might dream about but most just won't, and if they want high power for
a bass they'd head towards getting an Ampeg capable of 300 W using 6 x
6550. I guess ppl already with an ampeg could plug in KT120,and maybe
they'd get 330Watts with the same load, and maybe more with a slightly
lower load but SUSTAINED power with a sine wave measured at clipping
might show the PSU limitations.

Probably, the KT120 will last longer in an amp which has just the one
shared grid bias voltage supply, or two, to the 1, 2, or 3 tubes each
side of the PP circuit, and this is because one would use the same
bias current in KT120 as in 6550, in a bass amp. But tube matching
drifts well apart as tubes age, and Pda for each tube of a six pack
might vary between say 40 Watts and 15 Watts, and soon, at least one
or two 6550 will overheat and run away thermally and die, maybe
causing collateral damage. But KT120 have 60W Pda limit, so large
variations in Pda with ageing will e better tolerated. Of course the
ONLY correct way to apply fixed bias is to have a separate adjustment
for EACH output tube no matter what that tube is, which means having 6
pots and 6 LEDs to tell you if the bias of any one tube is too high.
But all such measures of intelligent management of errant tube
behaviour in the music instrument amp setting have been universally
condemned by all amp makers - too fukkin difficult, fidly, and too
fukkin expensive to put in during construction, and if we did we'd be
rooned. The major brands of amps treat the buying public as dumb jerks
who are NOT intelligent, and who exist to be relieved of their cash in
the easiest legal way possible.

Probably, one could build a nice powerful amp using compact sized
forced air cooled transmitting triodes or tetrodes normally used in
high power RF transmitters. May as well use mosfets, eh?

> would work great. Then again so would some pentodes or tetrodes, the
> 4-65A is much underused as is the 4E27. The 813 is too physically big.-

There is no end to what's possible, but there are severe limitations
on what is profitable, and what will sell like hot cakes to the
gullible and fickle public. Marketing is a bastard of the thing;
caused a few suicides.....

Patrick Turner.

Dau Bine

unread,
Nov 21, 2021, 3:34:29 PM11/21/21
to
Hey. There is 650V max voltage limit recommendation for KT120 in triode mode. So how about a p-p KT120 triode strapped, supplied at 600V into Ra-a 10k load, please? Thanks.

Big Bad Bombastic Bob

unread,
Nov 23, 2021, 5:12:29 PM11/23/21
to
On 11/21/21 12:34, Dau Bine wrote:
> Hey. There is 650V max voltage limit recommendation for KT120 in triode mode. So how about a p-p KT120 triode strapped, supplied at 600V into Ra-a 10k load, please? Thanks.
>

are you looking for a design to build? single-ended triode config?

I think you would be better with a push-pull triode config if you want
triode final stage. And I'd suggest sticking with the reference designs
when possible. You should be able to find some online... maybe for a
different output tube but then you make little tweeks here and there for
output impedence and bias and whatnot, and voila!

also consider that when an output transformer voltage swings positive on
the plate, you can nearly double B+ voltge at the plate.

Or were you actually looking at a resistive plate load? (not recommended).

You can get inexpensive output transformers from guitar repair parts
places online - I think 'amplifiedparts.com" is one of them.


--
(aka 'Bombastic Bob' in case you wondered)

'Feeling with my fingers, and thinking with my brain' - me

'your story is so touching, but it sounds just like a lie'
"Straighten up and fly right"

Dau Bine

unread,
Nov 23, 2021, 5:42:50 PM11/23/21
to
Hey. Thanks for the answer. I.m pretty much aware by You recommendation but this is what I have on my shelves. A kt120 matched quad, a hefty power transformer able to provide B+ 600v/1.5A and a Hammond 1650R 5k Ra-a. And I wonder if such of load is able to provide decent results for a push pull project with 4xKt120 in triode mode please. Thanks.

Big Bad Bombastic Bob

unread,
Nov 24, 2021, 1:29:34 PM11/24/21
to
oh ok I think I ge4t it. I assume it's a standard 5k CT since you ar
doing triode mode. For 600V B+ at 2.5k it's about 200ma total for full
swing (100V on plate). I think a pair of KT120s in triode mode might be
able to do that. So now you just need to bias it properly. Make sure
the output transformer has sufficient power handling though, since a lot
of current could saturate the core and cause clipping and serious plate
overheating.

The data sheet is here:
https://www.thetubestore.com/lib/thetubestore/TS-KT120.pdf

According to that the triode config can go up to 230ma for a single
tube. Using a quad means less likely to fry them.

The 5k transformer would probably get you ~100W (50WRMS). So if you
wanted to max them out (as a quad) you should get a 2.5k CT output
transformer that can handle that much power. Then current will be as
high as 400ma across 1.25k (appx 500V) which is 200W (100W RMS).

you'll need a somewhat beefy driver/splitter stage - the spec says max
51k on the G1 (fixed bias) and I'm guessing you'll bias it to at least
-100V. So that's a pretty good signal (200V peak-peak) going in, driven
into a (possible) 25k load impedence.

I'd suggest something like 2 12AU7s to drive the control grids, a common
cathode amplifier plus direct coupled cathode follower for each.

12AU7 has max cathode current of 22ma. With a swing of +/- 10ma driving
a 25k load, that's 250V (more than enough signal). So thenit's just a
matter of living within the limits of the 200V heater/cathode voltage
limit (this is always a problem with direct-coupled tubes and phase
splitters). Depending on the design, you could make this work with
direct coupling the 12AU7b cathode to G1 on the KT120s. Then your fixed
bias suppply will have to tolerate up to 40ma of current swing,
regardless.

And so on.
Message has been deleted
Message has been deleted

Dau Bine

unread,
Nov 28, 2021, 7:15:37 PM11/28/21
to
Thanks You for the answer. The 1650R 5k Ra-a is rated at 316 mA per side. Also I take in consideration to buy a kt88 quad to use it with this OT and to drive it from ecc99 plate for gain consideration. Do you think it work please? Thanks.

Big Bad Bombastic Bob

unread,
Nov 29, 2021, 2:58:44 PM11/29/21
to
well, "work" most likely. You have to select the output transformer's
impedance based on the expected voltage/current swing of the output
tube. So you use the max/min current of the linear range on the
performance chart, figure out what G1 volts needs to be biased at, etc.
based on all of that. So in short at max current plate volts will be a
bit under 100V at the max current, and quiescent (AB mode) will be high
enough to not have crossover distortion of any kind [so is still in the
linear range of the G1:plate curve]. That's typically how the designers
would pick the transformer for a reference design, anyway.

I did not see any plate volt/current curves (with G1 voltage references)
in the docs I found, but if you look for them you should be able to find
them.

If you want to limit distortion you could try biasing the output tubes
as class A. You get higher total current through the transformer
though, so you will need to use one that is beefy enough. Class A push
pull should get you very low distortion open-ended, at the expense of
gain and power requirements.

[this is why I use NFB insead]

G.Re

unread,
Dec 3, 2021, 3:44:51 AM12/3/21
to
I did not see any plate volt/current curves (with G1 voltage references)
in the docs I found, but if you look for them you should be able to find
them.

Here's a KT120 datasheet with Ia/Ua and Ia/Ug1 curves in pentode mode with
Ug2 reference . Also a Ia/Ua curve set in triode mode.
The published curves seem to originate from (an) individual tube(s) rather
than being nominal curves.
https://frank.pocnet.net/sheets/084/k/KT120.pdf

Gio

Dau Bine

unread,
Dec 5, 2021, 11:22:47 AM12/5/21
to
Thank You Gents!
Well, have not to much options over there : a hammond 1650R Ra-a 5k with 300 mA current ability, and a hefty 1650W Ra-a 1.9k , over 800 mA. I think I may use the W with its 4 ohm output for 8 ohm load to get Ra-a 3.8K for a KT quad at 600V

Big Bad Bombastic Bob

unread,
Dec 9, 2021, 12:40:51 PM12/9/21
to
there's NO kill like OVERkill like NO kill I know... (that monstrous
transformer used that way)

Dau Bine

unread,
Dec 9, 2021, 5:24:17 PM12/9/21
to
...I know, almost ridiculous for 88 W rms I got it...

Dau Bine

unread,
Dec 10, 2021, 11:19:41 PM12/10/21
to
Fortunately, smallest 1650R works as well making also a ridiculous situation using a kt quad for only 60w around, but it sound better than mismatched 1650w btw.

Dau Bine

unread,
Dec 14, 2021, 10:43:55 PM12/14/21
to
And I have a winner for my project. 4xkt88 svetlana into hammond 1650R for Ra-a 2.5k ( 16 ohm tap was used with 8 ohm load) at 600v. Got exactly 100W rms with the tubes biased 60 percent. Tested down to 40cps at nominal power and no sign of partial saturation. Screen grids under 6w dissipation. Very happy with result. Thanks for helping!
0 new messages